delayed_jobs save completed jobs

Question

Is there a parameter I can pass to delayed_job that will prevent it from deleting completed jobs from the delayed_jobs table?

kind of like the destroy_failed_jobs but for completed jobs

any ideas?

Answer 1

Expanding on @jefflunt's answer.

I added a migration to create a table to contain the completed jobs

class CreateCompletedJobs < ActiveRecord::Migration
  def change
    create_table :completed_jobs do |t|
      t.integer  "priority", :default => 0
      t.integer  "attempts", :default => 0
      t.text     "handler", :limit => 2147483647
      t.datetime "run_at"
      t.datetime "completed_at"
      t.string   "queue"
      t.timestamps
    end
  end
end

Then a module

class CompletedJob < ActiveRecord::Base
end

Finally added the hook to the job I want to store

  def success job
    save_completed_job job
  end

  private

  def save_completed_job job
    CompletedJob.create({
      priority: job.priority,
      attempts: job.attempts,
      handler: job.handler,
      run_at: job.run_at,
      completed_at: DateTime.now,
      queue: job.queue
    })
  end

Since I have more then one job I placed the success method hook in a module and included it in all the jobs I would like to store. (Note: some aren't worth storing)

Answer 2

It doesn't appear so. From the README: https://github.com/tobi/delayed_job

By default, it will delete failed jobs (and it always deletes successful jobs). If you want to keep failed jobs, set Delayed::Job.destroy_failed_jobs = false. The failed jobs will be marked with non-null failed_at.

You would probably need to hook into its destroy method such that it copies the job to another, separate table that keeps the list of completed jobs for you, or to simply log which jobs have been done to a file, if a log is all you need.

What you don't want is to leave the jobs in the delayed_jobs table, for a couple of reasons. First, because delayed_jobs uses that table as it's TODO list - you want it to only be things that still need to be done. Second, if you hacked it to keep all jobs in the same table, then the delayed_jobs table only grows, which would slow down the processing of delayed_jobs over time, as the query to find jobs that have not yet been completed would have to filter out those that have .

Answer 3

You could use something like acts_as_paranoid .

It basically works like this:

• #destroy is overwritten so that records are not deleted from the database but instead a deleted_at timestamp is set
• a default scope is installed so that .find (and all its variations) don't return items from the database where the deleted_at timestamp is set

Here is an introduction: http://technicalpickles.com/posts/embracing-your-paranoia-with-acts_as_paranoid/

Beware: acts_as_paranoid is very old and there are maybe better solutions, so check out https://www.ruby-toolbox.com/categories/Active_Record_Soft_Delete